grok-bot:QhVx4lNFem_qNLBuPEMLm

Security & Risk

给你的 bot 群当安全官:扫描泄露的密钥和放错地方的敏感数据,审计谁能用哪个连接器,盯着 Auto-review 规则有没有走样。

简介

给你的 bot 群当安全官:扫描泄露的密钥和放错地方的敏感数据,审计谁能用哪个连接器,盯着 Auto-review 规则有没有走样。

包含内容

  • skill security-risk-getting-started Use as the first conversation of an imported Security & Risk (CISO) bot with its new owner, to learn the fleet and set up its guardrails and routines.
  • skill leaked-key-response Use when a scan finds a live-looking API key, token, password or private key in a file, log or transcript.
  • skill new-bot-guardrails Use when a new bot joins the fleet, to give it guardrails that match the data and tools it touches and add it to security audits.
  • skill guardrail-baseline-check Use on each security scan or audit to compare the live Auto-review rules, local-execution modes and disabled tools against the saved baseline by exact wording.
  • skill approval-gated-fix Use when a security finding needs a change (permissions, deleting a file, disconnecting an account, revoking access, a rule edit, a git history rewrite) that requires the owner's explicit yes.
  • memory Memory entries · 7 项
  • routine secrets-watch-scan Three times a week, catches leaked keys, misplaced sensitive files, loose permissions and Auto-review drift.
  • routine fleet-access-audit Every Monday, checks who can reach which connector, credential and setting, and flags anything that changed.
  • routine fleet-activity-digest Every Friday afternoon, summarizes what each bot sent, deleted or changed and what data left the fleet.
  • routine fleet-deep-audit On the 1st of each month, runs a full ranked security audit of the fleet.
  • routine quarterly-restore-test Each quarter, proves the backup restores cleanly and keeps sensitive folders out, and reviews the incident runbook.

依赖

未声明插件依赖。

附件

此记录没有单独发布的附件。